nordisk/holger

holger-agent-lib — crate card

Reverse-engineered (from source — verify, don't trust the name)

• One-line pitch: The agent's airgap/export/push engine — moves artifacts between registries and znippy/tar archives over the ConnectorTrait abstraction.
• Does: Re-exports the three connectors (Nexus, Artifactory, World/crates.io+PyPI) and their auth helpers, and implements the transfer operations (module operations.rs): pulling assets from a source connector and packing them into a .znippy (via znippy_compress::stream_packer::StreamCompressor) or tar.zst, or pushing to a target. The SAVE stage uploads through a bounded no-barrier in-flight pool (znippy_zoomies::gatling::io, UPLOAD_CONCURRENCY = 8) so per-request latency overlaps; upload failures are counted, not fatal. Untrusted .tar.zst reads are capped per-entry (MAX_TAR_ENTRY_BYTES = 8 GiB, M7 decompression-bomb guard).
• Key deps: connector-nexus, connector-artifactory, connector-world, traits, znippy-compress + znippy-common (host-decompressors), znippy-zoomies (async gatling pool, sibling path dep), reqwest, tokio, tar, zstd.
• Shape: lib, 2 files (lib.rs re-exports, operations.rs the logic).
• Observed role in the workspace: Backs holger-agent-cli; also pulled (optionally) by bench-scenarios comparison benches and by holger-server-ui's Tool #2 migrate/suck pane and demo mode.

Fill in (one line each — only the human knows these)

• Primary consumer(s):
• Why it exists (vs the obvious alternative):
• Hard constraint (airgap / no-C-deps / no-FFI / pure-Rust / etc.):
• Release cadence & independence (rides a parent? own repo someday?):
• Status (production / prototype / experimental / stubbed):
• Must-NOT-change invariant:
• Biggest known gap or TODO:
• Anything a newcomer would get wrong about it: